Representation of chemical structures

Abstract

At the root of applications for substructure and similarity searching, reaction retrieval, synthesis planning, drug discovery, and physicochemical property prediction is the need for a machine‐readable representation of a structure. Systematic nomenclature is unsuitable, and notations and fragment codes have been superseded, except in certain specific applications. Connection tables are widely used, but there is no formal standard. Recently the International Union of Pure and Applied Chemistry (IUPAC) International Chemical Identifier (InChI) has started to attract interest. This review also summarizes the representation of chemical reactions and three‐dimensional structures. © 2011 John Wiley & Sons, Ltd. WIREs Comput Mol Sci 2011 1 557–579 DOI: 10.1002/wcms.36This article is categorized under: Computer and Information Science > Chemoinformatics

Many InChIs and quite some feat

“…So the InChI project seeks to convert the structure, in the form of its connection table [11] to a unique string of characters by fixed algorithms, generating the InChI using non-proprietary, open source software. The identifier is unique: the same label always means the same substance, and the same substance always receives the same label.…”

A possible extension to the RInChI as a means of providing machine readable process data

“…Projects are being undertaken to extend the representation to reactions and polymers, which is facilitated by the fact that due to its hierarchical nature new layers can be added relatively easily [16, 35]. …”

“…All layers aside from the main one are optional, and will only appear if the corresponding information has been provided in the source file [16, 36]. If the same structure has been drawn at two different levels of detail, the InChI for the one with less detail forms a subset of the one with more [15].…”

Silver threads

“…Work continues on the special problems of structure representation: more unusual forms of stereochemistry, polymers, and generic structures, for example [5]. There was much research into the handling of generic structures (so-called Markush structures) in the 1980s, culminating in the launch of two proprietary systems: Thomson Reuters' Markush DARC and CAS' MARPAT [6].…”

“…There has been considerable interest in this field of late. Text mining [7][8][9][10][11][12][13] recognizes chemical entities in journal articles or patents and extracts them for conversion to connection tables, and, in some cases, into IUPAC International Chemical Identifiers (InChIs) [5]. Related work converts images of chemical structures into connection tables [14][15][16][17][18][19].…”